Multi-Camera Fiducial Marker Detection in Dynamic Lighting

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current fiducial marker detection systems face challenges in reliable and high-speed detection, especially in environments with varying light intensities and dynamic scenes, such as outer space or industrial sites, where existing image sensors struggle with limited dynamic range, leading to incomplete data capture and inefficient processing.

Innovation Solution

The implementation of a multi-camera array system with different optical filters and the use of advanced marker detection algorithms, including the Auxiliary Tracking Algorithm and Image Sectioning Algorithm, which leverage previous image frames to improve processing efficiency and handle larger images with limited processing power, and the detection of markers in high dynamic range situations by combining images with varying exposures.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a single image sensor is used for marker detection, then the device complexity is low, but the reliability of detection in high dynamic range situations deteriorates due to limited dynamic range

Engineering Contradiction:
Improvedetection reliabilityVSAvoiddevice complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent divides the detection task across multiple image sensors, each specialized for specific lighting conditions. The system segments the detection function by using different sensors with different optical filters to capture different portions of the light spectrum or different exposure levels, then combines these segmented results to achieve reliable detection across high dynamic range situations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the parameters of the image sensing system by using multiple sensors with different optical filters (wavelength sensitivity, exposure times, aperture settings). This allows the system to capture images under varying lighting conditions and combine them to achieve reliable marker detection across high dynamic range environments.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If conventional marker detection algorithms are used, then the processing is simple, but the productivity deteriorates due to inefficient processing of large images

Engineering Contradiction:
Improveprocessing efficiencyVSAvoidalgorithm complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the image processing task by dividing large images into smaller regions or blocks that can be processed independently and in parallel. This segmentation approach maintains simplicity while improving processing efficiency by reducing the computational burden on single processors and enabling distributed or parallel processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary actions by pre-processing images to identify potential marker locations or regions of interest before applying the full detection algorithm. This preliminary step reduces the amount of data that needs to be processed in detail, thereby improving overall processing efficiency without requiring overly complex algorithms.

Inventive Principle:
Principle #10Preliminary action

3Loss of time

If image processing is performed on each frame independently, then the algorithm simplicity is maintained, but the loss of time increases due to redundant processing of similar frames

Engineering Contradiction:
Improveprocessing timeVSAvoidalgorithm complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The patent performs preliminary actions by using information from previously processed frames to guide the processing of current frames. This may include using detected markers from previous frames as starting points for current detection, or using motion compensation techniques to reduce redundant processing. This approach significantly reduces processing time while maintaining algorithmic coherence.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent maintains continuity of useful action by carrying forward detection results, intermediate calculations, or processed data from one frame to the next. This allows the system to build upon previous work rather than starting from scratch for each frame, thereby reducing overall processing time while maintaining a systematic approach.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentEP3146729B1System comprising a helmet, a multi-camera array and an ad hoc arrangement of fiducial marker patterns and their automatic detection in images
Publication Date: 2024.10.16 MILLENNIUM THREE TECHNOLOGIES INC
  • EP3146729B1 patent drawingFigure 1
  • EP3146729B1 patent drawingFigure 2~3
  • EP3146729B1 patent drawingFigure 4~5

AI summary

Fiducial markers are printed patterns detected by algorithms in imagery from image sensors for applications such as automated processes and augmented reality graphics. The present invention sets forth extensions and improvements to detection technology to achieve improved performance, and discloses applications of fiducial markers including multi-camera systems, remote control devices, augmented reality applications for mobile devices, helmet tracking, and weather stations.